A Reversed Order And Tolerance Method Based On Fuzzy Grade For Multi-objective Decesion

This paper focuses on the model building and the solving method of the multi-objective method which is based on the fuzzy rank, then coming up with evaluation indicator and building another model which meets decision makers' changeable preference.Because in the real multi-objective programming problems, decision makers tend to offer an unclear preference order among these goals, can not offer a strict order as the goal programming needs. This paper bases on the priority of the goal programming, studying on the existing solving methods firstly, and then improving the traditional methods, from the thought of tolerance step by step, extending the previous rank concept, innovating the priority structure, to form a fuzzy preference structure. This method tolerates the objects which obeys the reverse thought, just begin with the lowest object and then to the up one, at the same time, defining different tolerance variates which are according to different constraint conditions. As a result, getting a satisfying answer. To evaluate the objects by defining evaluation indicator as satisfying degree and optimizing degeree, this thought makes decision makers' evaluation basing on the numerical which makes the evaluation more objective. According to different evaluating results, building different correction models, which can meet decision makers' changeable preference.Introducing examples to do simulation testing.In goal programming, some problems may have no real satisfactory answers, and anther problems may need to give the weight coefficient number subjectively. However, the method in this article not only makes sure the objectivity of the solving process, but also obtain the real satisfactory answer. Comparing with the existing new methods, from the method based on alternative tolerance and based on genetic algorithm, and shows that the method in this article can optimize the solving speed and solution quality. In the cases of decision makers change the preference order, this method can do modification to models and get the real satisfactoryanswers. Making the method of this article combined with enterprise capital structure decision-making practice, and according to the narrow definition of capital structure, do a discussion between debt capital and equity capital when do the capital structure decision-making. And at the same time, determine its composition.The simulation result shows that the method in this article can get satisfactory answer easily and solve the multi-objective problems.